Method and device for determining fetal prenatal diagnosis threshold

Abstract

The invention provides a method and a device for determining a fetal prenatal diagnosis threshold value. The method comprises the following steps: obtaining effective data volumes and GC contents of a negative sample and a positive sample; constructing a reference database by using the m negative samples to obtain an average value mean and a variance sd of the reading number of the m negative samples in the specific region; preparing a mixed sample by using the n negative samples and the i positive samples to obtain n * i * 10 mixed samples; calculating the fetal concentration ff of the n negative samples, the i positive samples and the n * i * 10 mixed samples and the average value x of the reading number in the specific area; calculating by utilizing the average value mean, the variance sd, the fetal concentration ff and the average value x to obtain a plus Z value of each sample; and obtaining a threshold value of the plZ value by using an ROC curve. By means of the method, false positive and false negative can be reduced, a threshold value with more accurate judgment performance is obtained, and the precision of noninvasive prenatal screening is improved.

Description

Technical Field

[0001] The present invention relates to the field of gene detection technology, and in particular to a method and device for determining a threshold value for prenatal diagnosis of a fetus. Background Art

[0002] Non-invasive prenatal screening (NIPS) uses molecular genetic technologies such as high-throughput sequencing to detect fetal free DNA fragments in maternal plasma during pregnancy to assess the risk of common chromosomal aneuploidy and other genomic diseases in the fetus.

[0003] Studies have shown that trisomy 21, 18, and 13 has an incidence rate of 0.2%, and the proportion of fetuses carrying clinically significant pCNV (pathogenic chromosome copy number variation) in pregnant women can reach 1.6%-1.7%, which is one of the important genetic causes of birth defects such as congenital malformations and intellectual disabilities. Therefore, prenatal screening and prenatal diagnosis of the fetus are of great significance for the prevention and control of birth defects.

[0004] At present, in the process of non-invasive prenatal screening, the threshold for distinguishing positive and negative is often fixed threshold method, which is commonly used to calculate the Z value of the area to be tested. If the absolute value of the Z value is greater than 3, the area is considered abnormal. However, the fixed threshold method does not take into account the effect of fetal concentration on the Z value, which is prone to false positives or false negatives. Another common method is to use a certain amount of positive samples and a certain amount of negative samples to determine the threshold through the ROC method. However, the incidence of each specific disease area is often low, and it is difficult for researchers to obtain enough positive samples to draw the ROC curve.

[0005] Therefore, how to provide a more accurate method for prenatal screening of whether the fetus carries the pathogenic pCNV and determine the positive and negative judgment threshold is of great significance for prenatal diagnosis and prevention of birth defects. Summary of the invention

[0006] The main purpose of the present invention is to provide a method and device for determining a prenatal diagnosis threshold for a fetus, so as to solve the problem of low accuracy of the positive and negative thresholds in the prior art for prenatal diagnosis of whether a fetus carries a pathogenic pCNV.

[0007] In order to achieve the above-mentioned object, according to a first aspect of the present invention, a method for determining a threshold for fetal prenatal diagnosis is provided, the method comprising: performing quality control on genome sequencing data of m+n negative samples and i positive samples to obtain effective data volume and GC content of negative samples and positive samples; constructing a reference database using m negative samples to obtain the mean value mean and variance sd of the number of read segments in a specific region of the m negative samples; preparing a mixed sample using n negative samples and i positive samples, and separating each positive sample with a data volume ratio of 1:1, 1:2, or 3:4 to that of the positive sample. 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, and mixed with n negative samples in turn to obtain n*i*10 mixed samples; calculate the fetal concentration ff and the average value x of the number of reads in a specific area of ​​n negative samples, i positive samples and n*i*10 mixed samples; use the average value mean, variance sd, fetal concentration ff and average value x to calculate the plusZ value of each sample; use the ROC curve to obtain the threshold of the plusZ value; where m is ≥1000, n is ≥200, and i is ≥10.

[0008] Furthermore, constructing a reference database using m negative samples includes: dividing the genome of each negative sample and the reference genome into a plurality of one-to-one corresponding windows of the same length respectively; counting the number of reads in each window in each negative sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and performing GC content correction on the number of reads in each window in each negative sample to obtain the number of reads in each window in each negative sample that have been corrected for GC content; calculating the ratio of the number of reads in each window in each negative sample that have been corrected for GC content to the total number of reads in all windows in each negative sample that have been corrected for GC content, taking the total data volume as 40M, performing data volume correction on the number of reads in each window in each negative sample that have been corrected for GC content according to the ratio to obtain the number of reads in each window in each negative sample that have been corrected for data volume; calculating the mean value mean and variance sd of the number of reads in all windows of the m negative samples in a specific area that have been corrected for data volume; preferably, the size of the window is 50-100kb, and the length of the overlapping part of every two adjacent windows is 0-50kb.

[0009] Furthermore, the algorithm seqFF was used to calculate the fetal concentration ff of n negative samples, i positive samples and n*i*10 mixed samples.

[0010] Furthermore, a method for calculating the average value x of the number of reads in a specific region of n negative samples, i positive samples and n*i*10 mixed samples includes: dividing each negative sample or positive sample and the reference genome into multiple one-to-one corresponding windows of the same length; counting the number of reads in each window in each negative sample or positive sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, correcting the number of reads in each window in each negative sample or positive sample for GC content, and obtaining the number of reads in each window in each negative sample or positive sample that has been corrected for GC content; calculating the ratio of the number of reads in each window in each negative sample or positive sample that has been corrected for GC content to the total number of reads in the window in each negative sample or positive sample that has been corrected for GC content, and taking the total data volume as 40M, performing GC correction on each negative sample according to the ratio. The number of reads in each window in the negative sample or the positive sample after GC content correction is corrected for data volume to obtain the number of reads in each window in each negative sample or the positive sample after data volume correction; the average value x of the number of reads in all windows of each negative sample or the positive sample in a specific area after data volume correction is calculated; the average value x of the number of reads in all windows of each positive sample in a specific area is added to the average value x of the number of reads in all windows of n negative samples in a specific area, and the data volume ratios of positive samples and negative samples are 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, respectively, to obtain the average value x of the number of reads in all windows of each mixed sample in a specific area; preferably, the size of the window is 50-100kb, and the length of the overlapping part of each 2 adjacent windows is 0-50kb.

[0011] Furthermore, the plusZ value is calculated using plusZ=(x-mean) / (sd*ff); preferably, the threshold of the plusZ value obtained using the ROC curve includes: recording the diagnosis result of the negative sample as negative, recording the diagnosis result of the positive sample and the mixed sample as positive, drawing the ROC curve, and obtaining the threshold of the plusZ value.

[0012] Furthermore, after obtaining the threshold of the plusZ value, the method further includes: statistically calculating the fetal concentrations and plusZ values ​​of negative samples, positive samples and mixed samples, and the fetal concentration at which the diagnostic results of ≥95% of the samples can be normally detected is the lowest fetal concentration.

[0013] In order to achieve the above-mentioned purpose, according to a second aspect of the present invention, a device for determining a threshold for fetal prenatal diagnosis is provided, the device comprising: a quality control module, configured to perform quality control on genome sequencing data of m+n negative samples and i positive samples, and obtain the effective data volume and GC content of negative samples and positive samples; a reference database construction module, configured to construct a reference database using m negative samples, and obtain the average mean and variance sd of the number of read segments of the m negative samples in a specific area; a mixed sample module, configured to prepare a mixed sample using n negative samples and i positive samples, and mix each positive sample with a data volume ratio of 1:1, 1:2, 1:1 to negative samples, and 1:2 to positive samples. :3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, are mixed with n negative samples in sequence to obtain n*i*10 mixed samples; the calculation module is set to calculate the fetal concentration ff and the average value x of the number of reads in a specific area of ​​n negative samples, i positive samples and n*i*10 mixed samples; the plusZ value calculation module is set to use the average value mean, variance sd, fetal concentration ff and average value x to calculate the plusZ value of each sample; the threshold determination module is set to use the ROC curve to obtain the threshold of the plusZ value; wherein, the value of m is ≥1000, the value of n is ≥200, and the value of i is ≥10.

[0014] Furthermore, the reference database construction module includes: a first window division unit, which is configured to divide the genome of each negative sample and the reference genome into a plurality of one-to-one corresponding windows of the same length; a first GC content correction unit, which is configured to count the number of reads in each window in each negative sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each negative sample to obtain the number of reads in each window in each negative sample that has been corrected for GC content; a first data volume correction unit, which is configured to calculate the number of reads in each window in each negative sample that has been corrected for GC content. The ratio of the number of content-corrected reads to the total number of reads corrected for GC content in all windows in each negative sample, taking the total data volume as 40M, performing data volume correction on the number of reads corrected for GC content in each window in each negative sample according to the ratio, and obtaining the number of reads corrected for data volume in each window in each negative sample; a first mean value mean and variance sd calculation unit is configured to calculate the mean value mean and variance sd of the number of reads in all windows of the m negative samples corrected for data volume in a specific area; preferably, the window size is 50-100kb, and the length of the overlapping part of each 2 adjacent windows is 0-50kb.

[0015] Furthermore, the calculation module includes: a fetal concentration ff calculation unit, which is configured to calculate the fetal concentrations ff of n negative samples, i positive samples and n*i*10 mixed samples using seqFF.

[0016] Furthermore, the calculation module also includes: an average value x calculation module; the average value x calculation module includes: a second window division unit, which is configured to divide each negative sample or positive sample and the reference genome into multiple one-to-one corresponding windows of the same length respectively; a second GC content correction, which is configured to count the number of reads in each window in each negative sample or positive sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each negative sample or positive sample to obtain the number of reads in each window in each negative sample or positive sample that has been corrected for GC content; a second data volume correction, which is configured to calculate the ratio of the number of reads in each window in each negative sample or positive sample that has been corrected for GC content to the total number of reads in the window in each negative sample or positive sample that has been corrected for GC content, and take the total data volume as 40M, perform GC content correction on each negative sample according to the ratio. The number of reads in each window in the negative or positive sample that has been corrected for GC content is corrected for data volume to obtain the number of reads in each window in each negative sample or positive sample that has been corrected for data volume; the second mean and variance sd calculation unit is configured to calculate the average value x of the number of reads in all windows of each negative sample or positive sample in a specific area that has been corrected for data volume; the average value x of the number of reads in all windows of each positive sample in a specific area is compared with the average value x of the number of reads in all windows of n negative samples in a specific area, with the data volume ratios of positive samples and negative samples being 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, to obtain the average value x of the number of reads in all windows of each mixed sample in a specific area; preferably, the size of the window is 50-100 kb, and the length of the overlapping part of each 2 adjacent windows is 0-50 kb.

[0017] Furthermore, the plusZ value calculation module includes: a formula calculation unit, which is configured to calculate the plusZ value using plusZ=(x-mean) / (sd*ff); preferably, the threshold determination module includes: a ROC curve drawing unit, which is configured to record the diagnosis results of negative samples as negative, and the diagnosis results of positive samples and mixed samples as positive, draw the ROC curve, and obtain the threshold of the plusZ value.

[0018] Furthermore, after the threshold determination module, the device also includes: a data statistics unit, which is configured to count the fetal concentrations and plusZ values ​​of negative samples, positive samples and mixed samples, and a minimum fetal concentration determination unit, which is configured to be the fetal concentration at which the diagnostic results of ≥95% of the samples can be normally detected, i.e., the minimum fetal concentration.

[0019] In order to achieve the above object, according to a third aspect of the present invention, a computer-readable storage medium is provided, the storage medium including a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the above method.

[0020] In order to achieve the above objective, according to a fourth aspect of the present invention, a processor is provided, the processor being used to run a program, wherein the program runs the above method.

[0021] By applying the technical solution of the present invention, in the process of non-invasive prenatal screening of whether the fetus carries pathogenic pCNV, a method of mixing positive samples with negative samples is adopted. Without relying on a large number of positive samples, the fetal concentration of the reference sample is simultaneously used, thereby reducing the occurrence of false positives and false negatives, and being able to obtain a more accurate threshold for determining performance, thereby improving the accuracy of non-invasive prenatal screening, and being more feasible and extensive in actual operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0023] Figure 1 A schematic diagram of the ROC curve in the first embodiment of the present application is shown. DETAILED DESCRIPTION

[0024] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0025] As mentioned in the background technology, in the existing non-invasive prenatal screening process, the methods for determining the threshold of positive and negative discrimination are mainly fixed threshold method and ROC curve method, but these two commonly used methods have certain defects, such as not considering the difference in fetal concentration, prone to false positive and negative, or the number of positive samples for some diseases is not enough to support the drawing of the curve. Therefore, this application intends to provide a more accurate method for determining the threshold.

[0026] In a first typical embodiment of the present application, a method for determining a threshold for prenatal diagnosis of a fetus is provided, the method comprising: performing quality control on genome sequencing data of m+n negative samples and i positive samples to obtain effective data volume and GC content of negative samples and positive samples; constructing a reference database using m negative samples to obtain an average value mean and variance sd of the number of read segments of the m negative samples in a specific region; preparing a mixed sample using n negative samples and i positive samples, and separating each positive sample with a data volume ratio of 1:1, 1:2, 1: 3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, and mixed with n negative samples in turn to obtain n*i*10 mixed samples; calculate the fetal concentration ff and the average value x of the number of reads in a specific area of ​​n negative samples, i positive samples and n*i*10 mixed samples; use the mean mean, variance sd, fetal concentration ff and average value x to calculate the plusZ value of each sample; use the ROC curve to obtain the threshold of the plusZ value; where m is ≥1000, n is ≥200, and i is ≥10.

[0027] By mixing positive samples with negative samples, the problem of requiring a large number of positive samples in the process of threshold determination can be effectively solved. At the same time, taking the fetal concentration into account in the process of threshold determination can obtain a more accurate standard for determining the positivity or negativity of the variation.

[0028] In order to obtain more accurate gene analysis results, it is necessary to perform quality control screening on the gene sequencing results. In a preferred embodiment, quality control includes: filtering the original sequence and aligning it with the human reference genome, removing duplicate sequences, and extracting unique aligned base sequences.

[0029] Any method capable of performing the above-mentioned quality control is applicable to the present application. In a preferred embodiment, the original sequence filtering is performed by using a quality control tool (Trimmomatic-0.39 software) to process the original data from the MGI-2000 platform, and the adapter sequence, N-containing bases, and bases with a sequencing quality less than 15 in the original measured base sequence are removed to generate a new fastq file.

[0030] Any method capable of performing the above-mentioned quality control is applicable to the present application. In a preferred embodiment, the human reference genome is aligned by using alignment software (BWAv0.7.17) to align the filtered base sequence to the human reference genome sequence (version hg19); the repetitive sequence is removed by using de-repetitive sequence software (Picard v2.20.3) to remove the repeated sequencing reads of the alignment result BAM file and generate a new BAM file; the unique aligned base sequence is extracted by using software (Samtools v1.2) to extract the unique aligned base sequence in the alignment file after de-repetition and generate the final BAM file. The final BAM file is counted to obtain the effective data volume and GC content of the sample.

[0031] In order to further obtain a more accurate threshold, it is necessary to first construct a reference database using negative samples, which can provide a stable and accurate benchmark for subsequent sample analysis and can effectively compare and interpret the test results of different positive and negative samples. In a preferred embodiment, constructing a reference database using m negative samples includes: dividing the genome of each negative sample and the reference genome (human reference genome sequence (version hg19)) into multiple one-to-one corresponding windows of the same length; counting the number of reads in each window in each negative sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, correcting the number of reads in each window in each negative sample for GC content, and obtaining the number of reads in each window in each negative sample after GC content correction; calculating each in each negative sample The ratio of the number of reads corrected for GC content in the window to the total number of reads corrected for GC content in all windows in each negative sample, taking the total data volume as 40M, performing data volume correction on the number of reads corrected for GC content in each window in each negative sample according to the ratio, and obtaining the number of reads corrected for data volume in each window in each negative sample; calculating the mean value mean and variance sd of the number of reads in all windows of the m negative samples corrected for data volume in a specific area; preferably, the window size is 50-100kb, and the length of the overlapping part of each 2 adjacent windows is 0-50kb.

[0032] Any calculation method that can obtain the fetal concentration in the sample is applicable to the present application. In a preferred embodiment, seqFF is used to calculate the fetal concentration ff of n negative samples, i positive samples and n*i*10 mixed samples. The fetal concentration is the fetal sample concentration contained in the peripheral blood of pregnant women.

[0033] In order to obtain the standard read differences between negative and positive samples and the constructed reference database, the sequencing data of positive and negative samples were subjected to the same GC correction and data volume correction as those used to construct the reference database. In addition, the number of reads of the mixed sample was calculated based on the ratio of the negative and positive samples to the mixed sample.

[0034] In a preferred embodiment, a method for calculating the average value x of the number of reads in a specific region of n negative samples, i positive samples and n*i*10 mixed samples includes: dividing each negative sample or positive sample and the reference genome into multiple one-to-one corresponding windows of the same length respectively; counting the number of reads in each window in each negative sample or positive sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and correcting the number of reads in each window in each negative sample or positive sample for GC content to obtain the number of reads in each window in each negative sample or positive sample that has been corrected for GC content; calculating the ratio of the number of reads in each window in each negative sample or positive sample that has been corrected for GC content to the total number of reads in the window in each negative sample or positive sample that has been corrected for GC content, and taking the total data volume as 40M, calculating the ratio according to the ratio. The number of reads in each window in each negative sample or positive sample that has been corrected for GC content is corrected for data volume to obtain the number of reads in each window in each negative sample or positive sample that has been corrected for data volume; the average value x of the number of reads in all windows of each negative sample or positive sample in a specific area that has been corrected for data volume is calculated; the average value x of the number of reads in all windows of each positive sample in a specific area is compared with the average value x of the number of reads in all windows of n negative samples in a specific area, and the data volume ratios of positive samples and negative samples are 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, respectively, to obtain the average value x of the number of reads in all windows of each mixed sample in a specific area; preferably, the size of the window is 50-100 kb, and the length of the overlapping part of each 2 adjacent windows is 0-50 kb.

[0035] In order to further obtain a threshold value with higher accuracy in determining positivity and positivity, in a preferred embodiment, the plusZ value is calculated using plusZ=(x-mean) / (sd*ff); preferably, the threshold value of the plusZ value obtained using the ROC curve includes: recording the diagnosis result of the negative sample as negative, and the diagnosis result of the positive sample and the mixed sample as positive, drawing the ROC curve, and obtaining the threshold value of the plusZ value.

[0036] In order to further improve the efficiency of prenatal screening, the fetal concentration that can detect the results is directly used for judgment. In a preferred embodiment, after obtaining the threshold of the plusZ value, the method also includes: counting the fetal concentrations and plusZ values ​​of negative samples, positive samples and mixed samples. The fetal concentration that can be normally detected in the diagnostic results of ≥95% of the samples is the lowest fetal concentration.

[0037] In a second typical embodiment of the present application, a device for determining a threshold for prenatal diagnosis of a fetus is provided, the device comprising: a quality control module, configured to perform quality control on genome sequencing data of m+n negative samples and i positive samples, and obtain effective data volume and GC content of negative samples and positive samples; a reference database construction module, configured to construct a reference database using m negative samples, and obtain an average value mean and variance sd of the number of read segments of the m negative samples in a specific region; a mixed sample module, configured to prepare a mixed sample using n negative samples and i positive samples, and to mix each positive sample with a data volume ratio of 1:1, 1:2, 1:3, or 1:4 to that of the negative samples. 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, are mixed with n negative samples in turn to obtain n*i*10 mixed samples; the calculation module is set to calculate the fetal concentration ff and the average value x of the number of reads in a specific area of ​​n negative samples, i positive samples and n*i*10 mixed samples; the plusZ value calculation module is set to use the average value mean, variance sd, fetal concentration ff and average value x to calculate the plusZ value of each sample; the threshold determination module is set to use the ROC curve to obtain the threshold of the plusZ value; wherein, the value of m is ≥1000, the value of n is ≥200, and the value of i is ≥10.

[0038] Mixing positive samples with negative samples through the mixing module can effectively solve the problem of requiring a large number of positive samples in the process of threshold determination. At the same time, the plusZ value calculation module takes the fetal concentration into account in the process of threshold determination, which can obtain a more accurate positivity determination standard for this variation.

[0039] In order to obtain more accurate gene analysis results, it is necessary to perform quality control screening on the gene sequencing results. In a preferred embodiment, the quality control module includes: an original sequence filtering unit, a human reference genome alignment unit, a duplicate sequence removal unit and a unique alignment base sequence extraction unit.

[0040] Any method capable of performing the above-mentioned quality control is applicable to the present application. In a preferred embodiment, the original sequence filtering unit is configured to use a quality control tool (Trimmomatic-0.39 software) to process the original offline data from the MGI-2000 platform, remove the adapter sequence, N-containing bases, and bases with a sequencing quality less than 15 in the original measured base sequence, and generate a new fastq file.

[0041] Any method capable of performing the above-mentioned quality control is applicable to the present application. In a preferred embodiment, the human reference genome alignment unit is configured to use alignment software (BWA v0.7.17) to align the filtered base sequence to the human reference genome sequence (version hg19); the repetitive sequence removal unit is configured to use the de-repetitive sequence software (Picard v2.20.3) to remove the repeated sequencing reads of the alignment result BAM file and generate a new BAM file; the unique alignment base sequence extraction unit is configured to use software (Samtools v1.2) to extract the unique alignment base sequence in the alignment file after de-repetition and generate the final BAM file. The final BAM file is counted to obtain the effective data volume and GC content of the sample.

[0042] In order to further obtain a more accurate threshold, it is necessary to first use negative samples to build a reference database, which can provide a stable and accurate benchmark for subsequent sample analysis and can effectively compare and interpret the test results of different positive and negative samples. In a preferred embodiment, the reference database construction module includes: a first window division unit, which is configured to divide the genome of each negative sample and the reference genome into multiple one-to-one corresponding windows of the same length; a first GC content correction unit, which is configured to count the number of reads in each window in each negative sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each negative sample to obtain the number of reads in each window in each negative sample that has been corrected for GC content; a first data volume correction unit, which is configured to calculate the number of reads in each window in each negative sample that has been corrected for GC content. The ratio of the number of reads corrected for GC content to the total number of reads corrected for GC content in all windows in each negative sample is used. The total data volume is 40M, and the number of reads corrected for GC content in each window in each negative sample is corrected for data volume according to the ratio to obtain the number of reads corrected for data volume in each window in each negative sample. The first mean value and variance sd calculation unit is configured to calculate the mean value mean and variance sd of the number of reads in all windows of the m negative samples in a specific area after data volume correction. Preferably, the window size is 50-100kb, and the length of the overlapping part of each 2 adjacent windows is 0-50kb.

[0043] Any calculation method that can obtain the fetal concentration in the sample is applicable to the present application. In a preferred embodiment, the calculation module includes: a fetal concentration ff calculation unit, which is configured to use seqFF to calculate the fetal concentration ff of n negative samples, i positive samples and n*i*10 mixed samples.

[0044] In order to obtain the standard read differences between negative and positive samples and the constructed reference database, the sequencing data of positive and negative samples were subjected to the same GC correction and data volume correction as those used to construct the reference database. In addition, the number of reads of the mixed sample was calculated based on the ratio of the negative and positive samples to the mixed sample.

[0045] In a preferred embodiment, the calculation module also includes: an average value x calculation module; the average value x calculation unit includes: a second window division unit, which is configured to divide each negative sample or positive sample and the reference genome into multiple one-to-one corresponding windows of the same length respectively; a second GC content correction, which is configured to count the number of reads in each window in each negative sample or positive sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each negative sample or positive sample to obtain the number of reads in each window in each negative sample or positive sample that have been corrected for GC content; a second data volume correction, which is configured to calculate the ratio of the number of reads in each window in each negative sample or positive sample that have been corrected for GC content to the total number of reads in the window in each negative sample or positive sample that have been corrected for GC content, and take the total data volume as 40M, perform GC content correction on each window according to the ratio. The number of reads corrected for GC content in each window in the negative sample or the positive sample is corrected for data volume to obtain the number of reads in each window in each negative sample or the positive sample after data volume correction; the second mean and variance sd calculation unit is configured to calculate the average value x of the number of reads in all windows of each negative sample or positive sample in a specific area after data volume correction; the average value x of the number of reads in all windows of each positive sample in a specific area is compared with the average value x of the number of reads in all windows of n negative samples in a specific area, with the data volume ratios of positive samples and negative samples being 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, to obtain the average value x of the number of reads in all windows of each mixed sample in a specific area; preferably, the size of the window is 50-100 kb, and the length of the overlapping part of each 2 adjacent windows is 0-50 kb.

[0046] In order to further obtain a threshold with higher accuracy in determining positivity and positivity, in a preferred embodiment, the plusZ value calculation module includes: a formula calculation unit, which is configured to use plusZ=(x-mean) / (sd*ff) to calculate the plusZ value; preferably, the threshold determination module includes: a ROC curve drawing unit, which is configured to use the ROC curve to obtain the threshold of the plusZ value, including: recording the diagnosis result of the negative sample as negative, and the diagnosis result of the positive sample and the mixed sample as positive, drawing the ROC curve, and obtaining the threshold of the plusZ value.

[0047] In order to further improve the efficiency of prenatal screening, the fetal concentration that can detect the result is directly used for judgment. In a preferred embodiment, after the threshold determination module, the device also includes: a data statistics unit, which is configured to count the fetal concentrations and plusZ values ​​of negative samples, positive samples and mixed samples, and a minimum fetal concentration determination unit, which is configured to determine that the fetal concentration that can be normally detected in the diagnostic results of ≥95% of the samples is the minimum fetal concentration.

[0048] In a third typical embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium includes a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the above method.

[0049] In a fourth typical implementation of the present application, a processor is provided, the processor being configured to run a program, wherein the program runs the above method.

[0050] The present application is further described in detail below in conjunction with specific embodiments. These embodiments should not be construed as limiting the scope of protection claimed in the present application.

[0051] Embodiment 1

[0052] The 1,250 Digeorge syndrome-negative samples and 16 Digeorge syndrome-positive samples used in this application were all commercial samples from Beijing Youxun Medical Laboratory.

[0053] 1. Quality control of negative and positive samples

[0054] 1. Original sequence filtering: Use the quality control tool (Trimmomatic-0.39 software) to process the original data from the MGI-2000 platform, remove the adapter sequence, N-containing bases, and bases with sequencing quality less than 15 in the original measured base sequence, and generate a new fastq file;

[0055] 2. Alignment, de-repetitive sequences, and extraction of unique aligned base sequences: Use alignment software (BWAv0.7.17) to align the filtered base sequences to the human reference genome sequence (version hg19); use de-repetitive sequence software (Picardv2.20.3) to remove the repeated sequencing reads in the alignment result BAM file and generate a new BAM file; use software (Samtools v1.2) to extract the unique aligned base sequences in the de-repetitive alignment file and generate the final BAM file. Statistic the final BAM file to obtain the effective data volume and GC content of all negative samples and all positive samples.

[0056] 2. Use 1,000 negative samples to build a reference database.

[0057] 1. Divide the genome of each negative sample and the reference genome into multiple windows of the same length, with a window size of 50kb;

[0058] 2. GC content correction. Count the number of reads in each window of each negative sample that can be aligned with the corresponding window of the reference genome and the GC content of the reads in the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window of each negative sample to obtain the number of reads in each window of each negative sample after GC correction;

[0059] 3. Data volume correction. Calculate the ratio of the number of reads in each window of each negative sample after GC content correction to the total number of reads in all windows of each negative sample after GC content correction. Take the total data volume as 40M, and perform data volume correction on the number of reads in each window of each negative sample after GC content correction according to the ratio;

[0060] 4. Calculate the mean and variance sd of the number of reads in all windows of the m negative samples in a specific area after data volume correction.

[0061] 3. Determine the threshold using mixed samples

[0062] In addition, 200 negative samples and 10 positive samples were used to calculate the threshold of plusZ value for positive and negative judgment.

[0063] 1. Mixed samples. Each of the 10 positive samples was mixed with 200 negative samples at a data volume ratio of 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9 and 1:10 to obtain 20,000 mixed samples;

[0064] 2. Calculation of fetal concentration. Use seqFF software to count the fetal concentration of each sample in 200 negative samples, 10 positive samples, and 20,000 mixed samples;

[0065] 3. GC content and data volume correction.

[0066] 3.1 Divide each negative sample in the 200 cases or each positive sample in the 10 cases into multiple one-to-one corresponding windows with the same length as the reference genome. The window size is 100 kb, and every two adjacent windows have an overlap of 50 kb;

[0067] 3.2 Count the number of reads that can be aligned with the corresponding window of the reference genome in each window of each negative sample or each positive sample and the GC content of the reads of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window of each negative sample or each positive sample to obtain the number of reads after GC correction in each window;

[0068] 3.3 Calculate the ratio of the number of reads in each window of each negative sample or each positive sample after GC content correction to the total number of reads in all windows of each negative sample or each positive sample after GC content correction. Take the total data volume as 40M, and perform data volume correction on the number of reads in each window of each negative sample or each positive sample after GC content correction according to the ratio;

[0069] 3.4 Calculate the average x of the number of reads of all windows in a specific region of each negative sample or each positive sample corrected for the amount of data;

[0070] 3.5 The average value x of the number of reads of all windows in a specific area of ​​each positive sample and the average value x of the number of reads of all windows in a specific area of ​​200 negative samples are calculated, with the ratio of positive samples to negative samples being 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, to obtain the average value x of the number of reads of all windows in a specific area of ​​each mixed sample;

[0071] 4. Calculate plusZ value. According to the mean value mean and variance sd of the number of reads in all windows of the 1000 negative samples in the reference database in a specific region and the mean value x of the number of reads in all windows of the 200 negative samples, 10 positive samples and 20,000 mixed samples and the fetal concentration ff, calculate the plusZ value of each sample in the 200 negative samples, 10 positive samples and 20,000 mixed samples, plusZ = (x-mean) / (sd*ff);

[0072] 5. ROC determines the threshold for positive and negative judgment. The diagnosis results of negative samples are recorded as negative, and the diagnosis results of positive samples and mixed samples are recorded as positive. The ROC curve is drawn (such as Figure 1 The threshold of plusZ value for positive and negative judgment was determined as -2.85. A value less than or equal to -2.85 was judged as Digeorge positive, and a value greater than -2.85 was judged as negative.

[0073] 4. Performance verification.

[0074] Another 50 negative samples and 6 positive samples were used to calculate their respective plusZ values ​​using methods 3.1-3.4. -2.85 was used as the positive-negative judgment threshold to verify the performance of the threshold, and the positive-negative consistency rate was 100%.

[0075] Comparative Example 1

[0076] The "1. Quality control of negative samples and positive samples" and "2. Building a reference database using 1000 negative samples" of this comparative example are the same as those of Example 1. The specific methods of other operations are as follows:

[0077] 3. Traditional Z-value threshold

[0078] In addition, the same 200 negative samples and 10 positive samples of Digeorge syndrome as those in "III. Determining the threshold using mixed samples" in Example 1 were used to calculate the positive and negative judgment threshold.

[0079] 1. GC content and data volume correction

[0080] 1.1 Divide each negative sample in the 200 cases or each positive sample in the 10 cases into multiple one-to-one corresponding windows with the same length as the reference genome. The window size is 100 kb, and every two adjacent windows have an overlap of 50 kb;

[0081] 1.2 Count the number of reads that can be aligned with the corresponding window of the reference genome in each window of each negative sample or each positive sample and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window of each negative sample or each positive sample to obtain the number of reads after GC correction in each window;

[0082] 1.3 Calculate the ratio of the number of reads in each window of each negative sample or each positive sample after GC content correction to the total number of reads in all windows of each negative sample or each positive sample after GC content correction. Take the total data volume as 40M, and perform data volume correction on the number of reads in each window of each negative sample or each positive sample after GC content correction according to the ratio;

[0083] 2. Determination of traditional Z-value threshold

[0084] Calculate the average number of reads in all windows of the area to be detected for each negative sample or each positive sample after data volume correction, subtract the average mean, and compare the square difference sd to obtain the traditional Z value, and use the ROC curve to determine the threshold value of -3.2.

[0085] 3. Performance Verification

[0086] The threshold of this comparative ratio was verified using the same 50 negative samples and 6 positive samples as in Example 1, and the positive and negative coincidence rates were 95% and 80%, respectively.

[0087] From the above description, it can be seen that the above-mentioned embodiments of the present invention achieve the following technical effects: in the process of non-invasive prenatal screening of whether the fetus carries pathogenic pCNV, a method of mixing positive samples with negative samples is adopted. Without relying on a large number of positive samples, the fetal concentration of the reference sample is also used to reduce the occurrence of false positives and false negatives, and a more accurate threshold for determining performance can be obtained, thereby improving the accuracy of non-invasive prenatal screening, and being more feasible and extensive in actual operation.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for determining a fetal prenatal diagnosis threshold, characterized in that: The method comprises: After quality control of the genome sequencing data of m+n negative samples and i positive samples, the effective data volume and GC content of the negative samples and the positive samples are obtained; Using the m negative samples to construct a reference database, and obtaining the average mean and variance sd of the number of read segments of the m negative samples in a specific area; Using n negative samples and i positive samples to prepare mixed samples, each positive sample is mixed with n negative samples in sequence at a data volume ratio of the positive sample to the negative sample of 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10 to obtain n*i*10 mixed samples; Calculate the fetal concentration ff of the n negative samples, the i positive samples and the n*i*10 mixed samples and the average value x of the number of reads in the specific area; The plusZ value of each sample is calculated using the mean value mean, the variance sd, the fetal concentration ff and the mean value x; Using the ROC curve to obtain the threshold of the plusZ value; Among them, the value of m is ≥1000, the value of n is ≥200, and the value of i is ≥10.

2. The method according to claim 1, characterized in that The constructing a reference database using the m negative samples comprises: Dividing the genome of each negative sample and the reference genome into a plurality of one-to-one corresponding windows with the same length respectively; Counting the number of reads in each window of each negative sample that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and correcting the number of reads in each window of each negative sample for GC content to obtain the number of reads in each window of each negative sample after GC content correction; Calculate the ratio of the number of GC content-corrected reads in each window of each negative sample to the total number of GC content-corrected reads in all windows of each negative sample, and take the total data volume as 40M, perform data volume correction on the number of GC content-corrected reads in each window of each negative sample according to the ratio to obtain the number of data volume-corrected reads in each window of each negative sample; Calculate the mean value mean and variance sd of the number of reads of all windows of the m negative samples in the specific region corrected by the data amount; Preferably, the size of the window is 50-100 kb, and the length of the overlapping part of every two adjacent windows is 0-50 kb.

3. The method according to claim 1, characterized in that The algorithm seqFF is used to calculate the fetal concentration ff of the n negative samples, the i positive samples and the n*i*10 mixed samples.

4. The method according to claim 1, characterized in that: The method for calculating the average value x of the number of reads of the n negative samples, the i positive samples and the n*i*10 mixed samples in the specific area includes: Divide each of the negative sample or the positive sample and the reference genome into a plurality of one-to-one corresponding windows with the same length respectively; Counting the number of reads in each window in each of the negative samples or the positive samples that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and correcting the number of reads in each window in each of the negative samples or the positive samples for GC content, to obtain the number of reads in each window in each of the negative samples or the positive samples after GC content correction; Calculate the ratio of the number of reads corrected for GC content in each window of each negative sample or the positive sample to the total number of reads corrected for GC content in the window of each negative sample or the positive sample, and take the total data volume as 40M, perform data volume correction on the number of reads corrected for GC content in each window of each negative sample or the positive sample according to the ratio to obtain the number of reads corrected for data volume in each window of each negative sample or the positive sample; Calculate an average x of the number of reads of each negative sample or the positive sample in all windows of the specific region corrected for the data amount; The average value x of the number of reads of all windows in the specific area of ​​each positive sample and the average value x of the number of reads of all windows in the specific area of ​​n negative samples are calculated, with the data volume ratio of the positive samples to the negative samples being 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, to obtain the average value x of the number of reads of all windows in the specific area of ​​each mixed sample; Preferably, the size of the window is 50-100 kb, and the length of the overlapping part of every two adjacent windows is 0-50 kb.

5. The method according to claim 1, characterized in that The plusZ value is calculated using plusZ=(x-mean) / (sd*ff); Preferably, the using of the ROC curve to obtain the threshold of the plusZ value includes: recording the diagnosis result of the negative sample as negative, recording the diagnosis results of the positive sample and the mixed sample as positive, drawing the ROC curve, and obtaining the threshold of the plusZ value.

6. The method according to claim 1, characterized in that After obtaining the threshold of the plusZ value, the method further includes: The fetal concentration and the plusZ value of the negative sample, the positive sample and the mixed sample are counted, The fetal concentration at which the diagnostic results of ≥95% of the samples can be detected normally is the lowest fetal concentration.

7. A device for determining a threshold for prenatal diagnosis of a fetus, characterized in that: The device comprises: A quality control module is configured to perform quality control on genome sequencing data of m+n negative samples and i positive samples to obtain effective data volume and GC content of the negative samples and the positive samples; A reference database construction module is configured to construct a reference database using the m negative samples to obtain an average value mean and a variance sd of the number of read segments of the m negative samples in a specific region; A sample mixing module is configured to prepare a mixed sample using n negative samples and i positive samples, and each positive sample is mixed with n negative samples in sequence at a data volume ratio of the positive sample to the negative sample of 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10 to obtain n*i*10 mixed samples; A calculation module is configured to calculate the fetal concentration ff of the n negative samples, the i positive samples and the n*i*10 mixed samples and the average value x of the number of reads in the specific area; A plusZ value calculation module is configured to calculate the plusZ value of each sample using the mean value mean, the variance sd, the fetal concentration ff and the mean value x; A threshold determination module is configured to obtain a threshold of the plusZ value using a ROC curve; Among them, the value of m is ≥1000, the value of n is ≥200, and the value of i is ≥10.

8. The device according to claim 7, characterized in that The reference database construction module includes: The first window division unit is configured to divide the genome of each negative sample and the reference genome into a plurality of one-to-one corresponding windows with the same length respectively; A first GC content correction unit is configured to count the number of reads in each window in each of the negative samples that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each of the negative samples to obtain the number of reads in each window in each of the negative samples that has been corrected for GC content; A first data amount correction unit is configured to calculate the ratio of the number of GC content-corrected reads in each window of each negative sample to the total number of GC content-corrected reads in all windows of each negative sample, and perform data amount correction on the number of GC content-corrected reads in each window of each negative sample according to the ratio, taking the total data amount as 40M, to obtain the number of data amount-corrected reads in each window of each negative sample; A first mean value and variance sd calculation unit is configured to calculate the mean value and variance sd of the number of read segments of the m negative samples corrected for the data amount in all windows of the specific area; Preferably, the size of the window is 50-100 kb, and the length of the overlapping part of every two adjacent windows is 0-50 kb.

9. The device according to claim 7, characterized in that The calculation module comprises: The fetal concentration ff calculation unit is configured to calculate the fetal concentrations ff of the n negative samples, the i positive samples and the n*i*10 mixed samples using seqFF.

10. The device according to claim 7, characterized in that The calculation module also includes: an average value x calculation module; The average value x calculation module includes: The second window division unit is configured to divide each of the negative sample or the positive sample and the reference genome into a plurality of one-to-one corresponding windows with the same length; A second GC content correction is configured to count the number of reads in each window in each of the negative samples or the positive samples that can be aligned with the reference genome and the GC content of the reference sequence of the corresponding window in the reference genome, and perform GC content correction on the number of reads in each window in each of the negative samples or the positive samples to obtain the number of reads in each window in each of the negative samples or the positive samples that has been corrected for GC content; The second data amount correction is configured to calculate the ratio of the number of GC content-corrected reads in each window in each of the negative samples or the positive samples to the total number of GC content-corrected reads in the window in each of the negative samples or the positive samples, and to perform data amount correction on the number of GC content-corrected reads in each window in each of the negative samples or the positive samples according to the ratio, so as to obtain the number of data amount-corrected reads in each window in each of the negative samples or the positive samples; A second mean value mean and variance sd calculation unit is configured to calculate an average value x of the number of read segments of each negative sample or the positive sample in all windows of the specific region corrected for the data amount; The average value x of the number of reads of all windows in the specific area of ​​each positive sample and the average value x of the number of reads of all windows in the specific area of ​​n negative samples are calculated, with the data volume ratio of the positive samples to the negative samples being 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, and 1:10, to obtain the average value x of the number of reads of all windows in the specific area of ​​each mixed sample; Preferably, the size of the window is 50-100 kb, and the length of the overlapping part of every two adjacent windows is 0-50 kb.

11. The device according to claim 7, characterized in that The plusZ value calculation module includes: A formula calculation unit is configured to calculate the plusZ value using plusZ=(x-mean) / (sd*ff); Preferably, the threshold determination module includes: The ROC curve drawing unit is configured to record the diagnosis result of the negative sample as negative, the diagnosis result of the positive sample and the mixed sample as positive, draw the ROC curve, and obtain the threshold value of the plusZ value.

12. The device according to claim 7, characterized in that After the threshold determination module, the device further includes: A data statistics unit is configured to count the fetal concentration and the plusZ value of the negative sample, the positive sample and the mixed sample, The minimum fetal concentration determination unit is set to the fetal concentration at which the diagnostic results of ≥95% of the samples can be normally detected, which is the minimum fetal concentration.

13. A computer-readable storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

14. A processor, characterized in that: The processor is configured to run a program, wherein the program runs the method according to any one of claims 1 to 6.

Images